Near natural color polarization imagery fusion approach

To overcome the shortcomings of current false-color mapping methods for polarization imagery, a near natural color image fusion approach is addressed. Firstly, by polarization information analysis, the obtained images of total intensity, degree of linear polarization, and phase of polarization are mapped into the intensity component, saturation component and hue component of HSI color space, respectively. The result is then transformed from HSI color space to RGB color space, thereby the primary false-color fused image is acquired. After that, the false-color fused image is converted into YIQ color space, and the chromatic characteristics are calculated from the Y-channel, I-channel and Q-channel. And then, the chromatic characteristics of the reference image are adopted to modify those of the false-color fused image, and the image of degree of linear polarization is used to modulate the resultant I-channel image so as to optimize its color contrast. Finally, the resultant image is transformed back into RGB color space, so the near natural color image fusion result is yielded. Experimental results have shown its validity and feasibility of the proposed method.